基于数据融合和改进MUGM(1,m,w)的导弹装备故障预测

针对现代导弹装备系统组成复杂、结构关系模糊、特征参数获取不完整和不确定,造成其故障预测实现困难的问题,借鉴数据融合技术和灰色预测理论,提出了一种基于数据融合和改进多因素新陈代谢不等时距加权灰色预测模型(improved multi variables metabolism unequal interval weighted grey model， IMUGM(1,m,w))的导弹装备故障预测方法。首先,通过引入加权因子w的方式建立多因素不等时距加权灰色预测模型(UGM(1,m,w)),再通过初始值改进、残差修正、新陈代谢思想相结合的方式对模型进行改进;然后以特定个体的历史监测数据为基准,计算同类产品和特定个体的相应预测值及其与特定个体性能退化数值的Euclid距离,并根据Euclid距离确定隶属度权值,基于加权思想建立特定个体的性能退化模型,最后结合实时监测数据依次更新性能退化数据、Euclid距离、隶属度权值和性能退化模型,实现导弹装备故障预测,实例仿真及分析验证了方法的有效性。

Missile equipment fault forecast based on data fusion and improved MUGM(1,m,w)

In order to overcome the difficulty of modern missile equipment fault forecast, which is induced by the complexity of system composition, fuzziness of configuration connection and incomplete and uncertaint character parameters, according to data fusion technique and grey forecast theory, a new forecast method based on data fusion and improved multi variables metabolism unequal interval weighted grey model (IMUGM(1,m,w)model) is proposed. Firstly, multi variables unequal interval weight grey model(MUGM(1,m,w)model) is built by introducing weight gene and optimized by initial value optimization, residual error correction and metabolism．Then the specific individual’s historical measure data are used as the benchmark, and the same kind of products and the specific individual’s corresponding forecast values are calculated using IMUGM(1,m,w)models. The Euclid distances are used to determine degree of membership, so the individual’s performance degradation model is built on the basis of the degree of membership weighted method. Finally, the measurement data, Euclid distances, degree of membership and performance degradation model are updated with real time measurement data. The proposed method is applied to fatigue crack growth data, and the experimental results validate the validity. The result of simulating practical missile equipment fault forecast and analysis validates the validity.